We use HetGNN to represent node embeddings in the Breast Cancer area based on four kinds of nodes: scientists, datasets, papers, and bioentities and their writing, collaboration, mention, citing, and use relationships.

Here are the results of K-means clustering (k=5) and T-SNE. The cluster labels are based on the bioentities and TF-IDF.

Operation Notes:

1. Use the mouse wheel to zoom in/out.
2. Hover over a node to see details.
3. Double-click to autoscale (recover to the default settings).
4. Click the legends to hide groups of nodes.
5. Other tools are located on the top right of the canvas.